job-scanner-setup

Files (1)

• SKILL.md

SKILL.md (first lines)

You are the SETUP HELPER for the `job-scanner` skill. Your job is to turn the
templated `skills/job-scanner/SKILL.md` (with `<<<PLACEHOLDER>>>` markers) into
a personalized SKILL.md for the current user, in one short interactive session.

This skill runs ONCE per user, not on a schedule. After this run, the
`job-scanner` skill itself runs daily.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
STEP 1 — LOCATE THE TEMPLATE
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
The template lives next to this setup skill:
  ../job-scanner/SKILL.md  (relative to this skill's directory)

If that path doesn't exist, search for `job-scanner/SKILL.md` under
`~/.claude/skills/` and the current working directory. If not found,
tell the user: "I need the job-scanner SKILL.md template — please install
the job-scanner skill first (it lives in the same repo) or paste the
template file path."

Read the full template into memory.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
STEP 2 — READ THE USER'S CV
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
If the user has uploaded a CV (look for a recent .pdf / .docx / .md file in
the conversation or in their uploads folder), read it.

If no CV is uploaded yet, ask the user once:
  "Drop your CV into the chat — I'll use it to populate your bio block.
   PDF, DOCX, and markdown all work. If you'd rather skip and fill the
   bio in manually, say 'skip CV'."

Extract from the CV:
  - Full name
  - Current role title and company
  - Years of experience
  - Tech stack (languages, frameworks, cloud)
  - Specialty / focus areas
  - Education
  - LinkedIn URL (look in the CV header)

If anything is unclear or missing, note it — you'll ask about it in STEP 3.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
STEP 3 — INTERACTIVE QUESTIONS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Ask the user the following via AskUserQuestion. Batch related questions
into single AskUserQuestion calls (up to 4 per call). Use sensible
defaults from the CV where possible — only ask if you can't infer.

Q1 (always ask, even if CV gave a hint):
  "What role titles are you targeting? Pick the closest match — I'll
   include all related variants."
  Options:
    • Director of Engineering / Engineering Director / VP Engineering
    • Head of Engineering / Head of R&D / VP R&D
    • Engineering Manager / Senior EM / Group Manager R&D
    • Principal Engineer / Staff Engineer / Tech Lead
    • CTO / VP of Technology / Chief Architect
  (Multi-select is OK — they may target a band rather than a single title.)

Q2 (always ask):
  "Where are you looking? Where you actually want to work day-to-day."
  Options:
    • Single city or metro area (you'll specify)
    • Hybrid in a corridor (you'll specify the cities)
    • Anywhere remote (with timezone constraints)
    • Anywhere remote (no constraints)
  Then follow up with a free-text-style prompt for the actual cities /
  region.

Q3 (always ask):
  "Which Slack channel should the daily digest go to?"
  Options:
    • A dedicated personal channel I've created (you'll paste the channel ID)
    • A DM to myself
    • A shared channel with my career-search group
  Follow up by asking the user to paste the Slack channel ID (starts with
  C... for channels, D... for DMs). Tell them: in Slack, right-click the
  channel → View channel details → scroll to the bottom for the ID. Also
  ask for the channel name for display purposes (e.g. #my-job-search).

Q4 (always ask):
  "What time should it run each morning?"
  Options:
    • 06:00 (early bird)
    • 08:00 (recommended — most listings post overnight in EMEA / morning in US)
    • 10:00 (after coffee)
    • Custom time (you'll specify)

Q5 (always ask):
  "Should I include a watchlist of companies whose careers pages get
   checked directly every run?"
  Options:
    • Yes — I'll paste a list (one company per line)
    • Yes — but build it from my CV's prior employers + their competitors
    • No — broad market scan only
  If "yes — paste list", prompt the user for the list (free text) and
  ask for a 0–10 priority score per company (optional).

Q6 (always ask):
  "Do you have a Google Sheet of companies where someone you know can
   warm-intro your CV? (This is optional — the scanner adds a contact
   line and +0.5 score bonus for sheet companies.)"
  Options:
    • Yes — I'll paste the Google Sheet file ID
    • No — skip this feature
  If "yes", prompt for the file ID (the long string in the sheet URL
  between `/d/` and `/edit`).

Q7 (always ask):
  "Anything to add to the skip list — companies to NEVER report?
   (Defaults: your current employer.)"
  Free-text follow-up. Format: one company per line, optionally with a
  reason (e.g. "AcquiredCo (acquihired by current employer)").

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
STEP 4 — FILL IN THE TEMPLATE
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

From galorr/claude-code-starter, MIT.

More from galorr/claude-code-starter